New publication on decision fusion

New publication on decision fusion

August 1, 2015

Decision_fusion_loewOur new publication in ISPRS is accepted: “Decision fusion and non-parametric classifiers for land use mapping using multi-temporal RapidEye data”.

This study addressed the classification of multi-temporal satellite data from RapidEye by considering different classifier algorithms and decision fusion. Four non-parametric classifier algorithms, decision tree (DT), random forest (RF), support vector machine (SVM), and multilayer perceptron (MLP), were applied to map crop types in various irrigated landscapes in Central Asia. A novel decision fusion strategy to combine the outputs of the classifiers was proposed. This approach is based on randomly selecting subsets of the input data-set and aggregating the probabilistic outputs of the base classifiers with another meta-classifier. During the decision fusion, the reliability of each base classifier algorithm was considered to exclude less reliable inputs at the class-basis. The spatial and temporal transferability of the classifiers was evaluated using data sets from four different agricultural landscapes with different spatial extents and from different years. A detailed accuracy assessment showed that none of the stand-alone classifiers was the single best performing. Despite the very good performance of the base classifiers, there was still up to 50% disagreement in the maps produce by the two single best classifiers, RF and SVM. The proposed fusion strategy, however, increased overall accuracies up to 6%. In addition, it was less sensitive to reduced training set sizes and produced more realistic land use maps with less speckle. The proposed fusion approach was better transferable to data sets from other years, i.e. resulted in higher accuracies for the investigated classes. The fusion approach is computationally efficient and appears well suited for mapping diverse crop categories based on sensors with a similar high repetition rate and spatial resolution like RapidEye, for instance the upcoming Sentinel-2 mission.

 

you may also like:

Online course on Agricultural Land Use Mapping

Online course on Agricultural Land Use Mapping

Last week our colleague Michael Thiel gave an online course on Agricultural Land Use Mapping at WASCALs Doctoral Studies Program on Climate Change and Agriculture, which is based on the IRP/IFRA Campus in Katibougou, Mali. The course content included theories on...

AgriSens DEMMIN 4.0 Field Day

AgriSens DEMMIN 4.0 Field Day

On the 4th of November the AgriSens field day took place in Kruckow, Mecklenburg-Western Pomerania. Together with regional farmers, consultants and students the eight project partner succesfully presented and discussed the use of remote sensing technologies in...

most recent news:

New paper on degrees of urbanization

New paper on degrees of urbanization

New paper titled „To be, or not to be ‘urban’? A multi-modal method for the differentiated measurement of the degree of urbanization” has been published by Taubenböck & Team. Abstract: Today, 56.6% of the world's population is urban and the trend is rising; in...

new publication on earth observation and socio-economy

new publication on earth observation and socio-economy

Hannes Taubenboeck and colleagues published the article "Fusing Earth observation and socioeconomic data to increase the transferability of large-scale urban land use classification." in Remote Sensing of Environment. From the abstract: "Monitoring and understanding...

new publication on coast line dynamics

new publication on coast line dynamics

Ronja Lappe, a recent EAGLE graduate, published her thesis as article in Remote Sensing: "State of the Vietnamese Coast—Assessing Three Decades (1986 to 2021) of Coastline Dynamics Using the Landsat Archive". From the abstract: "Vietnam’s 3260 km coastline is densely...