Share Email Print
cover

Proceedings Paper

Development of a rule-based algorithm for rice cultivation mapping using Landsat 8 time series
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

In the framework of ERMES project (FP7 66983), an algorithm for mapping rice cultivation extents using mediumhigh resolution satellite data was developed. ERMES (An Earth obseRvation Model based RicE information Service) aims to develop a prototype of downstream service for rice yield modelling based on a combination of Earth Observation and in situ data. The algorithm was designed as a set of rules applied on a time series of Landsat 8 images, acquired throughout the rice cultivation season of 2014 from the plain of Thessaloniki, Greece. The rules rely on the use of spectral indices, such as the Normalized Difference Vegetation Index (NDVI), the Normalized Difference Water Index (NDWI), and the Normalized Seasonal Wetness Index (NSWI), extracted from the Landsat 8 dataset. The algorithm is subdivided into two phases: a) a hard classification phase, resulting in a binary map (rice/no-rice), where pixels are judged according to their performance in all the images of the time series, while index thresholds were defined after a trial and error approach; b) a soft classification phase, resulting in a fuzzy map, by assigning scores to the pixels which passed (as ‘rice’) the first phase. Finally, a user-defined threshold of the fuzzy score will discriminate rice from no-rice pixels in the output map. The algorithm was tested in a subset of Thessaloniki plain against a set of selected field data. The results indicated an overall accuracy of the algorithm higher than 97%. The algorithm was also applied in a study are in Spain (Valencia) and a preliminary test indicated a similar performance, i.e. about 98%. Currently, the algorithm is being modified, so as to map rice extents early in the cultivation season (by the end of June), with a view to contribute more substantially to the rice yield prediction service of ERMES. Both algorithm modes (late and early) are planned to be tested in extra Mediterranean study areas, in Greece, Italy, and Spain.

Paper Details

Date Published: 19 June 2015
PDF: 9 pages
Proc. SPIE 9535, Third International Conference on Remote Sensing and Geoinformation of the Environment (RSCy2015), 95350K (19 June 2015); doi: 10.1117/12.2193162
Show Author Affiliations
Christos G. Karydas, Aristotle Univ. of Thessaloniki (Greece)
Pericles Toukiloglou, Aristotle Univ. of Thessaloniki (Greece)
Chara Minakou, Aristotle Univ. of Thessaloniki (Greece)
Ioannis Z. Gitas, Aristotle Univ. of Thessaloniki (Greece)


Published in SPIE Proceedings Vol. 9535:
Third International Conference on Remote Sensing and Geoinformation of the Environment (RSCy2015)
Diofantos G. Hadjimitsis; Kyriacos Themistocleous; Silas Michaelides; Giorgos Papadavid, Editor(s)

© SPIE. Terms of Use
Back to Top