First-Class Land Cover with SPOT 6
Creating a map with land coverage on a regional/national scale without prior information and ground measurements to use, is a big challenge for both civil institutions (water management agencies, regional council) as well as private companies (insurance companies,…).
A classification can contain an important and very detailed number of classes, going from water to dense urban areas, through vegetation types, but can also be restricted to a certain type of land use.
When classifying agriculture and type of cultures, the information provided becomes of great interest to evaluate and control acreage, estimate yields; which implies access of multi-temporal data series to discriminate vegetation types.
Within the frame of the IGIS project*, aiming at setting up a National Spatial Data Infrastructure and a Remote Sensing Centre for the Republic of Serbia, SPOT 6 has been triggered to demonstrate its capacity to face this challenge, and provide an in-depth agricultural land cover.
Along one year, multiple SPOT 6 images have been acquired, post-processed and analysed to discriminate crops variety. This classification of living elements can only be achieved by monitoring plants growth profiles and cycles, thanks to biophysical parameters analysis, such as green cover fraction.
SPOT 6 spectral resolutions (1.5m Pan sharpen and 6m multispectral) and 12-bit image coding are assets to reach a high-level accuracy in a very fragmented agricultural landscape, such as in Serbia, and increase the quality of the segmentation stage.
- Access to intra-field information even for small parcels <0.05ha
- Reactive tasking : acquisitions at the right period of the cultural cycle
- Blue band: improved discrimination of the crops
- Textural information in addition to the multispectral analysis
- More accurate characterization of the cultural conditions (biophysical parameters extracted from the images)
In addition, the tasking capacity and the efficient management of cloud coverage, thanks to the 6 tasking plans per day, are essential for agriculture monitoring as periods for acquisitions are usually very narrow.
*This work has been carried out thanks to the cooperation between the Serbian Republic Geodetic Authority (RGA) and the French consortium IGN France International-Airbus Defence and Space.