The use of object oriented image analysis automatic extraction of unpaved roads in updating road maps: a lesson for the department of the Surveyor general

Abstract

This paper exhibits a method for extraction of unpaved roads from high resolution Pleiadis satellite imagery using Object Oriented Image Analysis (OOIA). The study area is Woodlands in Gweru Zimbabwe. Using ENVI EX, a rule based classification was performed. Classification rules were made from unpaved roads properties, by utilizing combination between unique multispectral bands, principal components and also utilizing combination of geometric attributes like width, length, neighbor and asymmetry and contextual ones. The outcome demonstrates that the OOIA algorithm effectively extracts unpaved roads from satellite images.