Bootstrapped CNNS for building segmentation on RGB-D aerial imagery

C. Sebastian, B. Boom, T. van Lankveld, E. Bondarev, P.H.N. de With

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

1 Citation (Scopus)
67 Downloads (Pure)

Abstract

Detection of buildings and other objects from aerial images has various applications in urban planning and map making. Automated building detection from aerial imagery is a challenging task, as it is prone to varying lighting conditions, shadows and occlusions. Convolutional Neural Networks (CNNs) are robust against some of these variations, although they fail to distinguish easy and difficult examples. We train a detection algorithm from RGB-D images to obtain a segmented mask by using the CNN architecture DenseNet. First, we improve the performance of the model by applying a statistical re-sampling technique called Bootstrapping and demonstrate that more informative examples are retained. Second, the proposed method outperforms the non-bootstrapped version by utilizing only one-sixth of the original training data and it obtains a precision-recall break-even of 95.10 % on our aerial imagery dataset.

Original languageEnglish
Title of host publicationISPRS TC IV Mid-term Symposium “3D Spatial Information Science – The Engine of Change”
EditorsS. Zlatanova, S. Dragicevic, G. Sithole
PublisherInternational Society of Photogrammetry and Remote Sensing (ISPRS)
Pages187-192
Number of pages6
DOIs
Publication statusPublished - 19 Sep 2018
Event2018 ISPRS TC IV Mid-Term Symposium on 3D Spatial Information Science - The Engine of Change - Delft, Netherlands
Duration: 1 Oct 20185 Oct 2018

Publication series

NameISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Number4
Volume4
ISSN (Print)2194-9042

Conference

Conference2018 ISPRS TC IV Mid-Term Symposium on 3D Spatial Information Science - The Engine of Change
CountryNetherlands
CityDelft
Period1/10/185/10/18

Keywords

  • aerial imagery
  • bootstrapping
  • building segmentation
  • deep learning

Fingerprint Dive into the research topics of 'Bootstrapped CNNS for building segmentation on RGB-D aerial imagery'. Together they form a unique fingerprint.

  • Cite this

    Sebastian, C., Boom, B., van Lankveld, T., Bondarev, E., & de With, P. H. N. (2018). Bootstrapped CNNS for building segmentation on RGB-D aerial imagery. In S. Zlatanova, S. Dragicevic, & G. Sithole (Eds.), ISPRS TC IV Mid-term Symposium “3D Spatial Information Science – The Engine of Change” (pp. 187-192). (ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences; Vol. 4, No. 4). International Society of Photogrammetry and Remote Sensing (ISPRS). https://doi.org/10.5194/isprs-annals-IV-4-187-2018