Master thesis

The main objective of the study is to create LULC classes in the Akshyirak massif, using OBIA between 1975 and 2014

  • Which LULC classes can be distinguished using OBIA and what are the accuracy levels achieved?
  • Which changes can be detected (type) and quantified (area) between considered LCLU classes in the Akshyirak massif from 1975 to 2014?
  • In particular, how much has the surface of the Petrov glacier and lake changed between 1975 and 2014?

Study area

study-areaPetrov glacier is located on the north-western slope of the Akshyirak massif, Kyrgyzstan. The glacier’s elevation reaches altitude 4700 m. In the foreground of it, there is a Petrov lake, in the past three decades, has doubled in size. The main reason for increasing the area of the lake is a melting of Petrov glacier.


Materials and methods

Landsat and Aster remote sensing imagery were used from 1975, 1990, 1999, 2005,2014 years. Pre-processing part consists of image georeferencing process of ASTER image, which did not contain coordinate reference system. The atmospheric noises on the data were removed through atmospheric correction. Spatial resolution of the Landsat OLI and ETM multispectral imagery was increased by applying image pan sharpening technique.

Two methods were applied in the classification process. The main task was to build rules with a certain threshold value for each band manually, namely a rule-based classification. The second method was sample-based nearest neighbour classifier, which classifies the image objects, according to their similarity to selected training samples.