Sentinel-1 time series data for crop identifications/discriminations
This work will be done in collaboration with IRRI. Prof. Andy Nelson and a local advisor from IRRI will also join the supervision team.
To inform policymakers about the agricultural landscapes and food production, accurate and timely monitoring of crop type is required. Multi-spectral satellite data such as Landsat, RapidEye and more recently Sentinel2 have been commonly used for crop-mapping. Nevertheless, the reliance of optical data on cloud cover limit image acquisitions at regular intervals, which is crucial for crop mapping. Spaceborne radar data can provide valuable information on vegetation cover stats irrespective of solar illumination and cloudiness. As such, Synthetic Aperture Radar (SAR) Sentinel-1 time series data due to its high spatial and temporal resolution offer a great opportunity to map crop types.
In the Philippines, a new project called PRIME - Pest Risk Identification and Management – is exploring the use of Sentinel-1 SAR imagery to generate maps of pest risk factors for rice. An extensive field survey has been conducted in several provinces to collect information about agricultural fields where the majority are planted with rice crop. The student will use time series Sentinel-1 imagery obtained over growing season and field data to identify and map crop types including (Rice, Maize, Mungbean, etc.) in the Philippines.
This study aims to detect/classify different crop categories and map their spatial distribution using time-series remote sensing data. Temporal analysis of Sentinel-1 SAR data of fields covering different crop types (Rice, Mazie, Mungbean, etc.) will be used to detect and map each crop category. Comprehensive field data, including crop types, field boundaries, planting and harvesting dates, soil and water management, will be available. These can be used to identify relationships between the SAR imagery and the recorded crop types using time series analysis and various classifications algorithms.