New publication on semi-supervised learning for classification of remote sensing image data

New publication on semi-supervised learning for classification of remote sensing image data

December 2, 2023

Researchers of the Earth Observation Center (EOC) of the German Aerospace Center (DLR), the Department of Geography of the University of Bonn, and our Earth Observation Research Cluster of the University Würzburg teamed up for a study on semi-supervised learning for classification of remote sensing image data. The paper titled “Semi-supervised learning with constrained virtual support vector machines for classification of remote sensing image data” was just published in the International Journal of Applied Earth Observation and Geoinformation by Christian Geiß, Patrick Aravena Pelizari, Ozan Tunçbilek and Hannes Taubenböck. This study has been conducted within the research project RIESGOS 2.0 (03G0905A-B), which is funded by the German Federal Ministry of Education and Research (BMBF).

 

Here is the abstract: We introduce two semi-supervised models for the classification of remote sensing image data. The models are built upon the framework of Virtual Support Vector Machines (VSVM). Generally, VSVM follow a two-step learning procedure: A Support Vector Machines (SVM) model is learned to determine and extract labeled samples that constitute the decision boundary with the maximum margin between thematic classes, i.e., the Support Vectors (SVs). The SVs govern the creation of so-called virtual samples. This is done by modifying, i.e., perturbing, the image features to which a decision boundary needs to be invariant. Subsequently, the classification model is learned for a second time by using the newly created virtual samples in addition to the SVs to eventually find a new optimal decision boundary. Here, we extend this concept by (i) integrating a constrained set of semilabeled samples when establishing the final model. Thereby, the model constrainment, i.e., the selection mechanism for including solely informative semi-labeled samples, is built upon a self-learning procedure composed of two active learning heuristics. Additionally, (ii) we consecutively deploy semi-labeled samples for the creation of semi-labeled virtual samples by modifying the image features of semi-labeled samples that have become semi-labeled SVs after an initial model run. We present experimental results from classifying two multispectral data sets with a sub-meter geometric resolution. The proposed semi-supervised VSVM models exhibit the most favorable performance compared to related SVM and VSVM-based approaches, as well as (semi-) supervised CNNs, in situations with a very limited amount of available prior knowledge, i.e., labeled samples.

 

Please find here the link to the full paper: https://www.sciencedirect.com/science/article/pii/S1569843223003953

 

This works adds to various earlier research studies focusing on the classification of remote sensing Imagery. Here are some examples:

 

 

you may also like:

Meeting with majors and local stakeholders within EO4CAM project

Meeting with majors and local stakeholders within EO4CAM project

On Wednesday, 30 October, a delegation from the EO4CAM project (comprising three members of the EORC and one from the ZKI, DLR) visited the small town of Markt Burgheim in Bavaria. They joined a meeting of local authorities and other stakeholders to discuss strategies...

social bouldering event

social bouldering event

Last Friday did we organize again a social event for all staff members and EAGLEs to get to know the new students, meet and chat with old students and just have a nice time outside the office for all staff members. We spend 4 hours together in the Rock In boulder gym...

snow cover conditions at Zugspitze with UAS sensors

snow cover conditions at Zugspitze with UAS sensors

Our UAS team travelled to Umweltforschungsstation Schneefernerhaus at Zugspitze to continue the time-series analysis for landscape analysis using drones. At an average elevation of 2500 meters, the team, Luisa Pflumm, Baturalp Arisoy, Konstantin Müller and Basil...

Guest lecture at KonGeoS conference

Guest lecture at KonGeoS conference

The KonGeoS conference (24-27.10.2024 in Würzburg) of geodesy student councils in Germany, Switzerland and Austria invited our Professor Hannes Taubenböck to give a guest lecture. On 25 October 2024, he gave a lecture on settlement development in Germany and discussed...

OPTIPLAN – Driving Innovation in Traffic and Urban Planning

OPTIPLAN – Driving Innovation in Traffic and Urban Planning

Driving Innovation in Traffic and Urban Planning: Insights from OPTIPLAN On October 24, 2024, the OPTIPLAN research consortium (German Aerospace Center (DLR)/University of Würzburg, Ingenieurbüro Behringer & Partner, and OBERMEYER GmbH) gathered to advance...