Calculation technique of satellite imaging quality assessment criterion based on geometrical concentration calculation

E. A. Maltsev, Yu A. Maglinets, R. V. Brezhnev

Research output: Contribution to journalArticlepeer-review

Abstract

The paper considers calculation methods of an objective satellite imaging quality assessment criterion based on geometrical concentration of image defects. Areas covered with clouds are considered as defects. An objective assessment implies an automatic calculation mode without the involvement of expert groups. Calculation of the geometrical concentration of objects on the plane based on Delaunay triangulation allows to proceed to the level of relational structures analysis, taking into account the information about mutual position of the objects in the image, and moreover to assess the nature of defects positioning in the form of a cloud cover. The paper shows the advantage of the proposed criterion of satellite image quality in comparison with the assessment based on the percentage of cloudiness. This criterion can be used in satellite data catalogs when selecting data for thematic processing. Approbation of calculation methods was conducted using random sampling of satellite images; the obtained quantitative results characterize the degree of images applicability for thematic processing. Recommendations for the application of the criterion under consideration in the selection and filtering of satellite images in the tasks of agricultural monitoring are formulated.

Original languageEnglish
Pages (from-to)35-44
Number of pages10
JournalSovremennye Problemy Distantsionnogo Zondirovaniya Zemli iz Kosmosa
Volume17
Issue number5
DOIs
Publication statusPublished - 2020

Keywords

  • Cloudiness
  • Geometrical concentration
  • Image processing
  • Image quality assessment
  • Remote sensing

Fingerprint

Dive into the research topics of 'Calculation technique of satellite imaging quality assessment criterion based on geometrical concentration calculation'. Together they form a unique fingerprint.

Cite this