Crop Mapping Improvement by Combination of Optical and SAR datasets

Abstract

Investigation of radar and optical data indices that contain a lot more information on landscapes and vegetation dynamics can be useful to identify opportunities and challenges in agricultural activities. In addition, the potential of synchronous implications of radar and optical data will be an effective method for agro-environmental monitoring and management to promote economic and environmental sustainability as monitoring programs. Crop discrimination as an agricultural monitoring system is a critical step regarding to estimate the area allocated to each crop type, computing statistics for crop control of area-based subsidies or crop production forecasting, environmental impact analysis and some other applications. Integrating both optical (reflectance) and Synthetic Aperture Radar (backscatter) multi-temporal features provides some advantages in terms of a more reliable crop map. We utilize multi-temporal Sentinel-1 Synthetic Aperture Radar (SAR) imagery and Sentinel-2 optical datasets in order to investigate the performance of the sensors backscatter and reflectance for temporal crop type mapping and the sustainable management of agricultural activities. Multi-temporal Sentinel-1, C-band VV and VH polarized SAR data and Sentinel2 optical data were acquired simultaneously by in-situ measurements for the study area. As preliminary results, it is concluded that the classification accuracies were improved results (5%) with using combinations of sensors. Classification accuracies of 93% were achieved in this study with integration use of SAR and optical data.