Black-Bridge data in the detection of forest area changes in the example of Sudety and Beskidy

Autorzy

• Tomasz Hycza Forest Research Institute, Department of Forest Resources Management, Sękocin Stary, Braci Leśnej 3, 05-090 Raszyn,
  Poland, phone: +48 22 7150343, e-mail: t.hycza@ibles.waw.
  
• Krzysztof Stereńczak Forest Research Institute, Department of Geomatics, Braci Leśnej 3, Sękocin Stary, 05-090 Raszyn, Poland
• Radomir Bałazy Forest Research Institute, Department of Geomatics, Braci Leśnej 3, Sękocin Stary, 05-090 Raszyn, Poland
  e-mail: radomir.balazy@gmail.com

Abstract

Two change detection techniques (NDVI differencing and post-classification analysis) were compared, in order to detect canopy cover changes in forests on the area of twelve forest districts in the Sudety and West Beskidy Mountains in Poland, using 2012 and 2013 Black-Bridge satellite images. Although the classification accuracy of the respective images was high (about 95%), the accuracy of the difference in bi-temporal images was much worse because of the short time between the dates of images and the imperfection of the algorithm calculating the unclear boundary between the forest and no-forest areas. NDVI differencing method and thresholding brought much better overall results, although roads, clouds and fogs caused much problem performing pseudo-changes.

Keywords

• Black-Bridge
• change detection
• post-classification analysis
• NDVI
• remote sensing