DEEP LEARNING FOR AUTOMATED MULTI-SCALE FUNCTIONAL FIELD BOUNDARIES EXTRACTION USING MULTI-DATE SENTINEL-2 AND PLANETSCOPE IMAGERY: CASE STUDY OF NETHERLANDS AND PAKISTAN

Speaker: Saba Zahid, Institute of Space Technology (IST) Islamabad, PakistanMonitoring crop health, ensuring food security, and implementing precision agriculture depend on accurate knowledge of agricultural field boundaries. Rapid temporal changes in agriculture field make challenge to abstract information from single date imagery in crop growing season. This task also become difficult in areas with small landholdings and lack of ground-based data for training of deep learning models. This study explores the effectiveness of multi-temporal satellite imagery for functional field boundary delineation using deep learning semantic segmentation architecture on two distinct geographical and multi-scale farming systems of Netherlands and Pakistan.