An Approach to Align Existing Vector Layers with High Resolution Satellite Images

Abstract

Geographical Information Systems (GIS) are highly used in various fields nowadays. So the accuracy of the geographical data is much needed. In GIS vector data is used to map the actual geographical objects because it is really hard to use actual geographical data all the time. When we analyzed the actual road layer data and its own vector layer we can see some erroneous deviation between those two. So in our research we are going propose a methodology to align the existing vector layers using more accurate high resolution satellite data. In our approach we replace the existing road vector layers with the new vector data generated by extracting actual road layer. Vector layers are generated by interconnecting co-ordinate points with small length line segments. In our approach first we filtered out the actual road segment using our own pixel based grouping process. Then the co-ordinate points are extracted by using customized Harris corner detection approach. Then we use this approach and extract some large number (15-20) of co-ordinate points and updated the existing vector layer by using the new vector data which are generated from extracted co-ordinates. Then the testing and evaluation is done under various conditions to check the accuracy level of our novel approach.