Application of Remote Sensing in Crop Monitoring

Remote sensing has become an important technology for monitoring agricultural crops and forecasting crop productivity over field, regional and national scales. Conventional crop monitoring depends heavily on field surveys, which are often time-consuming, labour-intensive and difficult to implement over large geographical areas. Remote sensing provides repeated, spatially explicit observations of crop condition through satellite, unmanned aerial vehicle (UAV) and airborne sensors. Multispectral, hyperspectral, thermal and synthetic aperture radar (SAR) observations can provide information on crop area, phenology, biomass, vegetation vigour, water status, nutrient stress, disease symptoms and other biophysical characteristics. Vegetation indices such as the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI) and red-edge indices have been widely used for crop monitoring. Recent developments in Sentinel-1, Sentinel-2, Landsat and other Earth-observation missions, together with machine learning and deep learning, have expanded the potential for near-real-time crop monitoring and yield forecasting. Recent reviews indicate a methodological shift from simple statistical relationships toward machine learning, deep learning, data assimilation and multi-sensor data-fusion approaches. However, remote-sensing-based forecasting remains affected by cloud cover, mixed pixels, limited ground observations, model transferability, crop diversity and uncertainty in yield measurements. This review discusses the principles, sensors, applications, forecasting methods, advantages, limitations and future prospects of remote sensing for agricultural crop monitoring and forecasting. Integration of remotely sensed observations with weather, soil, management and crop-growth information is likely to be particularly important for developing reliable operational agricultural decision-support systems.