Comparative analysis of crop maps for chosen test areas on the territory of Bulgaria and Romania using simulated Proba-V and Spot vegetation data
Eugenia Roumenina, Lachezar Filchev, Vassil Vassilev, Petar Dimitrov, Georgi Jelev, Gheorghe Stancalie, Elena Savin, and Denis Mihailescu
The operational capabilities of low-resolution satellite instruments, such as SPOT VEGETATION, MODIS, and NOAA AVHRR, make real-time and near-real time Earth Observation (EO) data indispensable for monitoring world vegetation resources. The data continuity of the SPOT VEGETATION mission is to be ensured by PROBA-V (“V” standing for “Vegetation”). The present study aims to reveal the potentialities of the PROBA-V mission and its advantages as compared to SPOT VEGETATION based on a comparison of accuracy assessments of crop identification maps. The scope of the study is achieved by using unsupervised ISODATA and supervised Maximum Likelihood Classification (MLC) algorithms on single-date PROBA-V simulated data and SPOT VEGETATION images, and a comparison of classification accuracies. The results show that single-date PROBA-V simulated images provide better crop map classifications than SPOT VEGETATION images with a 15 to 22% higher overall accuracy, and Kappa statistics of 0.31 to 0.33 higher for the Zhiten test area (Bulgaria) and the Fundulea test area (Romania), respectively. It was also established that single-date PROBA-V simulated data can serve as a basis for reliable crop identification of winter crops with high users and producers accuracy.
Submitted: 8 June 2012
Revised: 27 Nov 2012
Accepted: 12 Dec 2012
Published: 21 Dec 2012
Responsible editor: Ioannis Manakos
Roumenina E, L Filchev, V Vassilev, P Dimitrov, G Jelev, G Stancalie, E Savin & D Mihailescu, 2012.
Comparative analysis of crop maps for chosen test areas on the territory of Bulgaria and Romania using simulated Proba-V and Spot vegetation data.
EARSeL eProceedings, 11(2): 155-160