ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume V-3-2020
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., V-3-2020, 787–794, 2020
https://doi.org/10.5194/isprs-annals-V-3-2020-787-2020
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., V-3-2020, 787–794, 2020
https://doi.org/10.5194/isprs-annals-V-3-2020-787-2020

  03 Aug 2020

03 Aug 2020

LEARNING FROM NOISY SAMPLES FOR MAN-MADE IMPERVIOUS SURFACE MAPPING

C. Qiu1, P. Gamba2, M. Schmitt1, and X. X. Zhu1,3 C. Qiu et al.
  • 1Signal Processing in Earth Observation, Technical University of Munich (TUM), Munich, Germany
  • 2Telecommunications and Remote Sensing Laboratory, University of Pavia, Italy
  • 3Remote Sensing Technology Institute (IMF), German Aerospace Center (DLR), Wessling, Germany

Keywords: Classification, Fully convolutional networks (FCNs), Impervious surface mapping, Noisy samples, Robust loss function, Sentinel-2

Abstract. Man-made impervious surfaces, indicating the human footprint on Earth, are an environmental concern because it leads to a chain of events that modifies urban air and water resources. To better map man-made impervious surfaces in any region of interest (ROI), we propose a framework for learning to map impervious areas in any ROIs from Sentinel-2 images with noisy reference data, using a pre-trained fully convolutional network (FCN). The FCN is first trained with reference data only available in Europe, which is able to provide reasonable mapping results even in areas outside of Europe. The proposed framework, aiming to achieve an improvement over the preliminary predictions for a specific ROI, consists of two steps: noisy training data pre-processing and model fine-tuning with robust loss functions. The framework is validated over four test areas located in different continents with a measurable improvement over several baseline results. It has been shown that a better impervious mapping result can be achieved through a simple fine-tuning with noisy training data, and label updating through robust loss functions allows to further enhance the performances. In addition, by analyzing and comparing the mapping results to baselines, it can be highlighted that the improvement is mainly coming from a decreased omission error. This study can also provide insights for similar tasks, such as large-scale land cover/land use classification when accurate reference data is not available for training.