Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Towards grape-vine management based on mapping of airborne hyperspectral images

Paper presented at 9th International Symposium on Mobile Mapping Technology, MMT 2015, 9-11 December 2015, Sydney, Australia
: Knauer, Uwe; Edwards, Everard; McGrath, Andrew; Lieff, Wolfgang; Hacker, Jorg M.; Seiffert, Udo

Volltext urn:nbn:de:0011-n-4108438 (10 MByte PDF)
MD5 Fingerprint: a79aaad586da4dd6011e35e4c0dd5063
Erstellt am: 30.9.2016

2015, 5 S.
International Symposium on Mobile Mapping Technology (MMT) <9, 2015, Sydney>
Vortrag, Elektronische Publikation
Fraunhofer IFF ()

In this paper a number of potential applications for airborne hyperspectral imaging in grape-vine management are presented. The hyperspectral images are analysed with respect to different important aspects of canopy composition and structure such as leaf nitrogen content, leaf area and chlorophyll content. The article covers the whole processing chain from image acquisition to the mapping of the results. An NDVI based image segmentation, partial least squares regression models and prototype-based unsupervised clustering are among the developed and applied algorithms. A dataset from a canopy management study at Langhorn Creek has been used to test the approach. Limitations of mapping between hyperspectral data and ground-based reference data due to incomplete or unprecise information about sampling positions are reduced by an iterative partial least squares regression with resampling from the original input data.