Funding Body: German Federal Ministry for Economic Affairs and Energy (BMWi)
Duration: 2021 to 2023
Budget: 170.000 Euro
Role: PI and main author of proposal
The goal of the DESTAM project is to develop innovative AI-based methods to condense time series of optical satellite images using radar-based satellite imagery. The advantage of radar imagery is its independence from weather conditions and cloud cover, allowing regular observation of the Earth’s surface. The disadvantage is the lower information content compared to multi-spectral optical imagery, which makes it difficult to derive detailed land surface parameters.
By means of the AI-based time series, a temporally dense and application-oriented monitoring of agricultural areas will be enabled. The increased information gain from improved time series is helpful to overcome the high field-specific variability of crop systems in smallholder agriculture. In particular, monitoring at key phenological times, such as the beginning of the growing or harvest seasons, requires dense time series that have often been lacking until now. The methods to be developed will make it possible to condense time series specifically at individual data points and to evaluate these condensed time series subsequently with regard to agricultural parameters. The procedures resulting from the project will be used for monitoring anomalies in African subsistence farming and will be exploited accordingly. Decision-makers in crisis prevention and insurance companies can use these data to assess productivity bottlenecks and market needs at an early stage.
Selected Publications: tbd