Universität zu Köln

Huett, Christoph and Waldhoff, Guido (2018). Multi-data approach for crop classification using multitemporal, dual-polarimetric TerraSAR-X data, and official geodata. Eur. J. Remote Sens., 51 (1). S. 62 - 75. FIRENZE: ASSOC ITALIANA TELERILEVAMENTO. ISSN 2279-7254

Full text not available from this repository.

Abstract

Crop distribution information is essential for tackling some challenges associated with providing food for a growing global population. This information has been successfully compiled using the Multi-Data Approach (MDA). However, the current implementation of the approach is based on optical remote sensing, which fails to deliver the relevant information under cloudy conditions. We therefore extend the MDA by using Land Use/Land Cover classifications derived from six multitemporal and dual-polarimetric TerraSAR-X stripmap images, which do not require cloud-free conditions. These classifications were then combined with auxiliary, official geodata (ATKIS and Physical Blocks (PB)) data to lower misclassification and provide an enhanced LULC map that includes further information about the annual crop classification. These final classifications showed an overall accuracy (OA) of 75% for seven crop-classes (maize, sugar beet, barley, wheat, rye, rapeseed, and potato). For potatoes, however, classification does not appear to be as consistently accurate, as could be shown from repeated comparisons with variations of training and validation fields. When the rye, wheat, and barley classes were merged into a winter cereals class, the resultant five crop-class classifications had a high OA of about 90%.

Item Type:	Journal Article
Creators:	Creators Email ORCID ORCID Put Code Huett, Christoph UNSPECIFIED UNSPECIFIED UNSPECIFIED Waldhoff, Guido UNSPECIFIED UNSPECIFIED UNSPECIFIED
URN:	urn:nbn:de:hbz:38-202899
DOI:	10.1080/22797254.2017.1401909
Journal or Publication Title:	Eur. J. Remote Sens.
Volume:	51
Number:	1
Page Range:	S. 62 - 75
Date:	2018
Publisher:	ASSOC ITALIANA TELERILEVAMENTO
Place of Publication:	FIRENZE
ISSN:	2279-7254
Language:	English
Faculty:	Unspecified
Divisions:	Unspecified
Subjects:	no entry
Uncontrolled Keywords:	Keywords Language TIME-SERIES; RANDOM FOREST; RADAR IMAGES; INFORMATION; COVER; BAND; PHENOLOGY; RICE; SAR; MODELS Multiple languages Remote Sensing Multiple languages
Refereed:	Yes
URI:	http://kups.ub.uni-koeln.de/id/eprint/20289