State of the art in enhancement of classification results: Blog12

Research output: Book/ReportReportCommunication

7 Downloads (Pure)

Abstract

Land cover maps (LCMs) generated by classification of DSM-based orthoimages still need improvements regarding the smoothing of lines and generalizing the content. Human-caused objects like buildings, roads, bridges, and walls should be resolved by points, lines, and other geometric elements. Buildings should be represented by closed polygons with straight, parallel, and orthogonal lines. The corner points (vertices) of the polygons should have georeferenced coordinates.
Various authors published research in this field. First, the works of Wang (2016), Tasar et al. (2018), Mousa et al. (2019), Höhle (2021), Kong et al. (2023), and Schuegraf et al. (2024) are discussed, all of which use exclusively aerial photographs.
Original languageEnglish
Number of pages3
Publication statusUnpublished - 27 Jun 2024

Keywords

  • land cover map
  • enhancement
  • regularization

Fingerprint

Dive into the research topics of 'State of the art in enhancement of classification results: Blog12'. Together they form a unique fingerprint.

Cite this