A remote sensing (RS) platform consisting of a remote-controlled aerial vehicle (RAV) can be used to monitor crop, environmental conditions, and productivity in agricultural areas. However, the current methods for the calibration of RAV-acquired images are cumbersome. Thus, a calibration method must be incorporated into RAV RS systems for practical and advanced applications. Here, we aimed to develop a standalone RAV RS-based calibration system without the need for calibration tarpaulins (tarps) by quantifying the sensor responses of a multispectral camera, which varies with light intensities. To develop the standalone RAV-based RS calibration system, we used a quadcopter with four propellers, with a rotor-to-rotor length of 46 cm and height of 25 cm. The quadcopter equipped with a multispectral camera with green, red, and near-infrared filters was used to acquire spectral images for formulating the RAV RS-based standardization system. To perform the calibration study process, libraries of sensor responses were constructed using pseudo-invariant tarps according to the light intensities to determine the relationship equations between the two factors. The calibrated images were then validated using the reflectance measured in crop fields. Finally, we evaluated the outcomes of the formulated RAV RS-based calibration system. The results of this study suggest that the standalone RAV RS system would be helpful in the processing of RAV RS-acquired images.

It is important to be able to predict the yield and monitor the growth conditions of crops in the field to increase productivity. One way to assess field-based geospatial crop productivity is by integrating a crop model with a remote-controlled aerial system (RAS). The objective of this study was to simulate spatiotemporal barley growth and yield based on the development of a crop-modeling system integrated with RAS-based remote sensing images. We performed field experiments to obtain ground truth data and RAS images of crop growth conditions and yields at Chonnam National University (CNU), Gwangju, South Korea in 2018, and at Gyeongsang National University (GNU), Jinju, South Gyeongsang, South Korea in 2018 and 2019. In model calibration, there was no significant difference (p = 0.12) between the simulated barley yields and measured yields, based on a two-sample t-test at CNU in 2018. In model validation, there was no significant difference between simulated yields and measured yields at p = 0.98 and 0.76, according to two-sample t-tests at GNU in 2018 and 2019, respectively. The remote sensing-integrated crop model accurately reproduced geospatial variations in barley yield and growth variables. The results demonstrate that the crop modeling approach is useful for monitoring at-field barley conditions.

In temperate monsoon areas, major constraints of soybean production in drained paddy fields are excess soil water during monsoon seasons. To further understand how agronomic practices and weather course affect the yield of soybeans, we conducted field experiments at Gwangju, Korea (35°10′ N, 126°53′ E) over three years (2018–2020). Double-crop soybeans were grown at two tillage systems (TS) [rotary tillage (RT), deep plowing followed by rotary tillage (DPRT)] and three sowing dates (SD) (June 10–15, June 25–30, and July 10–15) in drained paddy fields. Flowering phenology (R2) was accelerated by 5 days with each 15-day delay in SD. This resulted in a significant reduction in vegetative growth up to R2, with subsequent reductions in CGR and NAR through R2–R5 (beginning of grain filling). With a 30-day delay in SD, yield was significantly reduced by 27.0%. The better performance of RT over DPRT was demonstrated by greater yields (13.7%). In addition, yield was greatly varied with weather volatility among years, ranging from 123.8 to 552.0 g m⁻². Weather volatility was the greatest contributor to yield variability (30.4%), followed by SD (17.0%) and TS (10.3%). Our results suggest that the yield might be mainly determined by how much growth has already been achieved before flowering and through R2–R5.

Deep learning (DL) and machine learning (ML) procedures are prevailing data-driven schemes capable of advancing crop-modelling practices that assimilate these techniques into a mathematical crop model. A DL or ML modelling scheme can effectively represent complicated algorithms. This study reports on an advanced fusion methodology for evaluating the leaf area index (LAI) of barley and wheat that employs remotely sensed information based on deep neural network (DNN) and ML regression approaches. We investigated the most appropriate ML regressors for exploring LAI estimations of barley and wheat through the relationships between the LAI values and four vegetation indices. After analysing ten ML regression models, we concluded that the gradient boost (GB) regressor most effectively estimated the LAI for both barley and wheat. Furthermore, the GB regressor outperformed the DNN regressor, with model efficiencies of 0.89 for barley and 0.45 for wheat. Additionally, we verified that it would be possible to simulate LAI using proximal and remote sensing data based on assimilating the DNN and ML regressors into a process-based mathematical crop model. In summary, we have demonstrated that if DNN and ML schemes are integrated into a crop model, they can facilitate crop growth and boost productivity monitoring.

Remote sensing technology has revolutionized agricultural monitoring by enabling real-time, large-scale crop growth assessment, facilitating precision agriculture practices, and supporting timely management decisions.^{1 – 3} For instance, satellite-derived vegetation indices, such as the normalized difference vegetation index (NDVI) and enhanced vegetation index, have been widely utilized to monitor crop growth conditions, estimate crop yields, and identify stress factors such as drought or nutrient deficiency across extensive agricultural landscapes.^{4 – 6} In addition, unmanned aerial vehicles equipped with multispectral or hyperspectral sensors have facilitated high-resolution monitoring of specific crop traits such as chlorophyll content, biomass, and leaf area index (LAI), leading to precise site-specific management practices.^7,8 Integrating remote sensing data into crop models has notably enhanced their accuracy and adaptability, thus improving the prediction of agricultural performance.^{9 – 11} These models effectively capture spatiotemporal variations in crucial physiological traits such as LAI, biomass accumulation, and yield, providing improved predictions crucial for optimal agricultural management.^12,13

This study developed a remote sensing-integrated crop model (RSCM) incorporating machine learning (ML) methodologies to predict maize yields across the US Corn Belt. The framework integrated MODIS vegetation indices with AgERA5 meteorological data to simulate leaf area index and biomass accumulation at 500 m resolution. Simulations across seven states (2012-2020) representing 70.7% of US maize cultivation achieved yield predictions ranging from 9.1±2.15 t ha-1 (South Dakota) to 11.7±1.96 t ha-1 (Iowa), with inter-regional variability of 0.2 to 2.6 t ha-1 reflecting diverse environmental conditions. The RSCM-ML approach successfully bridges traditional crop modeling and remote sensing integration, providing standardized geospatial datasets with broad applicability to diverse crop systems. The resulting standardized geospatial dataset is openly available via the National Research Data Platform and can support applications in crop monitoring, yield forecasting, and agroclimatic impact assessment. Future improvements through higher-resolution field observations and ensemble modeling will further enhance prediction accuracy for sustainable agricultural monitoring.