Accuracy assessment and classification efficiency of object-based image analysis of aerial imagery

Abstract

The paper focuses on the territorial differentiation of socioeconomic development of Poland between the years 2002-2014 and on geographic patterns of this differentiation according to the subregions ('podregiony' in Polish, NUTS 3 level). Eight partial indicators entering the composite indicator and also the average base index are applied. The analysis of the socioeconomic development of the subregions along the directional east-west gradient, rural-urban concentric gradients (around big cities) and the zones of subregions along the border of Poland with the surrounding countries are used to explain the observed differentiation. Polish subregions have undergone considerable development between 2002 and 2014, but the territorial differentiation of their development has changed only partially. The big Polish cities and also their suburban subregions have the best position of all; the worst are still the rural subregions of eastern Poland and the inner peripheries of Poland. The directional east-west gradient, the rural-urban concentric gradients, as well as the higher development of subregions at German, Czech and sea borders were confirmed.

Integrated Airborne LiDAR Data and Imagery for Suburban Land Cover Classification Using Machine Learning Methods

“…In a suburban land cover environment, the infrastructure is inadequate and the architectural style is not uniform, such as the cement roads are often mixed with bare soil, the greenland planning is confusing and the residential buildings with different heights, styles and materials. Additionally, some other factors such as the training sample datasets quality, classification methods and the design of classification scheme, affect the quality of classification outcomes [47]. These problems have hindered the accuracy of suburban land cover classification, which is an important reason why the OA accuracy of this paper has not reached more than 90%.…”