The use of radar images for detecting when grass is harvested and thereby improve grassland yield estimates

Bijeenkomst: EGF2015
Auteur: Grant K., Wagner M., Siegmund R. and Hartmann S.
ISBN: 978-9090-289-61-8
Jaar van uitgifte: 2015
Producttype: Paper
Cutting date and frequency are important parameters which, together with weather, soil conditions, botanical composition and fertilizer, determine grassland yields. However, cost- and time-efficient methods for recording cutting dates of grassland are currently lacking. Therefore, we developed a method for detecting cutting dates using changes in radar images of the sward surface. The combination of this method with a grassland yield model will result in more reliable and region-wide data on grassland yield estimates. For determining when grassland has been cut, robust amplitude-change detection techniques were used evaluating the amplitude or backscatter statistics before and after the cutting events in a test area in Germany. All detected cuts were verified according to in situ measurements and recorded in a GIS database. This method will be further adjusted to Sentinel-1 data which will then enable an area-wide and cost-efficient cutting-date detection service. The cutting frequency and yield data gained by this method are essential for optimising the use of grassland, for yield adjusted fertilisation, and the assessment of unused potential of grassland for alternative energies (e.g. biogas, solid fuel).